Metaphor Interpretation as Embodied Simulation

RAYMOND W. GIBBS, JR.

Abstract: Cognitive theories of metaphor understanding are typically described in

terms of the mappings between different kinds of abstract, schematic, disembodied
knowledge. My claim in this paper is that part of our ability to make sense of metaphorical
language, both individual utterances and extended narratives, resides in the automatic
construction of a simulation whereby we imagine performing the bodily actions referred
to in the language. Thus, understanding metaphorical expressions like ‘grasp a concept’
or ‘get over’ an emotion involve simulating what it must be like to engage in these
specific activities, even though these actions are, strictly speaking, impossible to physically
perform. This process of building a simulation, one that is fundamentally embodied in
being constrained by past and present bodily experiences, has specific consequences for
how verbal metaphors are understood, and how cognitive scientists, more generally,
characterize the nature of metaphorical language and thought.

1. Introduction

Understanding metaphor, similarly to all aspects of language, requires listeners to

draw inferences about speakers’ possible communicative intentions in saying what
they did. Although listeners can infer meanings that deviate from what speakers
strictly intend to convey, making an estimate of what speakers had in mind when
talking is a critical part of linguistic interpretation. One way of inferring what
speakers mean is to simulate what it must be like to be that person and have the
particular thoughts he or she had at the moment of production. My claim in this
paper is that imaginative simulation processes, which are fundamentally part of the
embodied mind, guide many aspects of metaphor understanding.
Consider some excerpts from an article, titled ‘Grief as a journey’ on a popular
internet website that provides information and support for professionals and
laypersons (http://www.helphorizon.com):

‘The loss of a loved one is not something that anyone ever ‘gets over’. We
may ‘get used to’ our loved one not being in our lives, but we never get over
the fact that a piece of our heart will be missing forever.
‘So, what to do if ‘getting over’ our loss is not a realistic option. We journey
through it, eventually recovering from our wounds. But just like other journeys
in our lives, we will need directions, supplies, plans, and most of all support
from others …. When we are in the throes of grief, we need to remember that
we’re not lost in the deepest depths of a dark cave with no way out. We are

Address for correspondence: Department of Psychology, University of California, Santa Cruz,

Santa Cruz, CA 95064, USA.
Email: gibbs@ucsc.edu
Mind & Language, Vol. 21 No. 3 June 2006, pp. 434–458.
© 2006 The Author
Journal compilation © 2006 Blackwell Publishing Ltd
 Metaphor Interpretation as Embodied Simulation 435

surely in the dark, but more like a dark tunnel. If we begin to move forward,
we will work our way through the tunnel to the other side—to where we can
begin to learn to live again.

None of the metaphorical phrases employed in these excerpts are especially novel
or poetic. Yet the various conventional expressions (e.g. ‘getting over’ something
in reference to an abstract entity) and elaborations of the basic grief as a journey
theme (e.g. needing ‘directions, supplies, plans’ etc.) nicely combine to form a
coherent scenario of the grief experience, one which most readers can readily
understand and, perhaps, appreciate. Given the conventional forms of metaphor
used here, one might assume, as do many metaphor scholars, that interpreting the
discourse is primarily a matter of looking up the encoded meanings of these
different phrases, maybe from a mental, phrasal lexicon, and then tying these
meanings together to create a well-formed mental model (e.g. a structured
propositional network) of the whole narrative.
However, my claim is that part of our ability to make sense of this narrative, and
its various conventional metaphors, resides in the automatic construction of a
simulation whereby we imagine performing the bodily actions referred to in these
excerpts. Thus, metaphorical references to getting over and through grief, needing
directions, supplies, plans and support from others in dealing with grief, being lost
in the dark tunnel of grief and moving forward to the other side, are all understood
by simulating what it must be like to perform these specific activities, even though
it is, strictly speaking, impossible to physically act on abstract entities like the
emotion of grief. We experience grief, in this case, as a process of moving through
‘affective space’ in which we imaginatively encounter different physical obstacles
and learn to overcome these in our ongoing emotional journey (Gibbs, 2006a).
This process of building a simulation, one that is fundamentally embodied in being
constrained by past and present bodily experiences, has specific consequences for
how verbal metaphors are understood, and how cognitive scientists, more generally,
characterize the nature of metaphorical language and thought.
I should note at the outset that the argument made here in favor of embodied
simulation is not intended as a comprehensive account of all verbal metaphor
understanding. The complexity of metaphoric language makes it unlikely, in my
view, that any single theory will be capable of explaining how verbal metaphors
come into being, and how they are ordinarily produced and interpreted. Yet many
aspects of metaphoric language appear to arise from, and continue to be grounded
in, patterns of embodied experience and may be understood via cognitive
simulations that are also fundamentally embodied. My aim in this paper is to
explore this claim and describe some linguistic and psycholinguistic evidence that
supports it.
At the same time, the idea that embodied simulation processes are important to
metaphor interpretation does not imply that these processes are unique to metaphor,
because psycholinguistic studies have recently shown that simulation is critical to
comprehending many types of nonmetaphorical language (see below). The
© 2006 The Author
Journal compilation © 2006 Blackwell Publishing Ltd
436 R. W. Gibbs

interesting point about metaphor understanding, however, is that people engage in

embodied simulations to interpret many verbal metaphors that refer to physically
impossible events, such as ‘moving forward through grief’. People’s easy
understanding of these metaphorical expressions through embodied simulations is
consistent with the idea that individuals readily conceive of many abstract topics in
embodied metaphorical ways.
Finally, my thesis that many kinds of metaphors are understood through
embodied simulations adopts a wide view of embodiment. Critical brain areas (e.g.
motor cortex) are likely recruited during ordinary linguistic processing of both
metaphorical and nonmetaphorical language. But as importantly, people’s intuitive,
felt, phenomenological experiences of their own bodies shape large portions of
metaphoric thoughtb and language use.

2. The Case for Embodied Metaphor

Making the case for embodied metaphor is the first step toward establishing my
claim that many verbal metaphors are specifically interpreted in terms of embodied
simulations. Simply put, one reason why people interpret many verbal metaphors
through embodied simulations is because this metaphoric language is rooted in
bodily processes that people may imaginatively recreate during their ordinary use
of such language. For instance, cognitive linguists have proposed that that many of
our concepts, including abstract ones, are grounded in, and structured by, various
patterns of our perceptual interactions, bodily actions, and manipulations of objects
(Gibbs, 1994, 2006a; Lakoff and Johnson, 1999). Recurring aspects of these
experiences, such as those associated with taking journeys, as manifested across
many sensory modalities and the kinesthetic body, give rise to ‘image schemas’
(e.g. SOURCE-PATH-GOAL) that are often mapped onto dissimilar domains to
create concrete understandings of abstract concepts. For example, people conceive
of emotional experiences like grief in terms of bodily actions performed upon
concrete entities and spaces (e.g. moving from a source along a path toward a
particular destination within the affective space associated with grief). In this
manner, bodily experiences provide the source domains for metaphorically
structuring aspects of abstract target domain spaces (e.g. the emotion of grief).
Systematic linguistic analyses of conventional and novel linguistic expressions (e.g.
‘moving through grief’, ‘getting over’ grief) across many spoken and signed
languages have revealed that a vast number of abstract ideas appear to be talked
about, and possibly understood, in terms of embodied metaphor (e.g. time,
causation, spatial orientation, political and mathematical ideas, emotions, the self,
concepts about cognition, morality) (Gibbs, 2006a; Lakoff and Johnson, 1999).
Embodied metaphor provides the grounding for many novel, creative metaphors
as well as it does conventional speech. Although some novel metaphors reflect
completely new source-to-target domain mappings, most novel metaphorical
expressions are creative instantiations of enduring conceptual metaphors (Gibbs,
© 2006 The Author
Journal compilation © 2006 Blackwell Publishing Ltd
 Metaphor Interpretation as Embodied Simulation 437

1994; Lakoff and Turner, 1989). One analysis of literary texts in Turkish shows,
for instance, that Turkish, similarly to many other languages, conceives of life as
movement along a path where birth is conceptualized as arrival, life is a journey
initiated by the arrival, and death is the departure that ends the journey (Ozcaliskan,
2003). Most generally, life is conceptualized as a region bounded by two doors,
one leading to life and the other to death, which gives rise to the following
conceptual mappings:

Source domain Target domain

Space bound by the two doors life
Being within the confines of a bounded space being alive
Physical threshold for change of location threshold for change of state
(e.g., doorway)
Crossing a physical threshold into a space being born
between two doors
Crossing a physical boundary to exit a space dying
between two doors
A moving entity human body

Two linguistic instantiations of these mappings are seen in the following translated
excerpts (Ozcaliskan, 2003, pp. 285 and 286):

‘The first moment I came into the world

I walked at the same time
In a caravanserai with two doors
I am going day and night’.
‘When the road of fate reaches the city of no beyond
A regretful hand will close the door of your life’.

Traveling along life’s journeys can take various forms, including, for example,
traveling though life as a sea voyage, as seen in the following poem:

‘I wandered here and there like many others

I wandered like dark rainy nights
Like a ship struggling on the ocean
I passed my days in unawareness’.

Linguistic analyses of literary texts, such as in the above, provide additional

evidence on the ubiquity of embodied metaphor in structuring the way people
talk and think about various abstract domains (Freeman, 1995). Psycholinguistics
studies have explored some of the implications of these ideas about embodied
conceptual metaphors to see if people actually use bodily experiences when
understanding different abstract concepts and the metaphorical language used to
© 2006 The Author
Journal compilation © 2006 Blackwell Publishing Ltd
438 R. W. Gibbs

talk of these concepts (Gibbs, 2006a). Again, my concern here, for the moment, is
to demonstrate that people recruit embodied metaphorical ideas in their creation
and understanding of many verbal metaphors. I will then expand on this argument
to suggest that the recruitment of embodied metaphors in some aspects of verbal
metaphor understanding is done imaginatively as people recreate what it must be
like to engage in similar actions. An important methodological element of the
research described in this section is the strategy to independently assess people’s
phenomenological intuitions about their bodily experiences and use this information
to make empirical predictions about individuals’ understanding of metaphorical
expressions.
For instance, one set of experiments investigated how people’s intuitions of the
bodily experience of containment, and several other image schemas, which serve
as the source domains for several important conceptual metaphors (e.g. ANGER
IS HEATED FLUID IN THE BODILY CONTAINER), underlie speakers’ use
and understanding of American idioms like ‘blow your stack’, and ‘flip your lid’
(Gibbs, 1992). These studies were designed to show that the specific meanings of
idioms arise from the source to target domain mappings of the conceptual
metaphors from which these expressions arise and maintain their currency in the
language. Most importantly, these metaphorical mappings preserve the cognitive
topology of their image-schematic source domains such as when the schema of
SOURCE-PATH-GOAL or CONTAINMENT is mapped onto the emotional
experience in phrases like ‘moving through grief’ or ‘blowing your stack’. Assessing
people’s intuitions about their embodied knowledge of source domains provides
evidence to make detailed predictions about what gets mapped onto different
target domains in metaphorical concepts, which in turn constrains people’s detailed
understandings of conventional metaphoric language referring to these concepts.
Participants in a first study were questioned about their understanding of events
corresponding to particular bodily experiences that were viewed as motivating
specific source domains in conceptual metaphors (e.g. the experience of one’s
body as a container filled with fluid from ANGER IS HEATED FLUID IN THE
BODILY CONTAINER). For instance, participants were instructed to imagine
the embodied experience of a sealed container filled with fluid, and then asked
something about causation (e.g. ‘What would cause the container to explode?’),
intentionality (e.g. ‘Does the container explode on purpose or does it explode
through no volition of its own?’), and manner (e.g. ‘Does the explosion of the
container occur in a gentle or a violent manner?’) of possible events within this
source domain. People were remarkably consistent in their responses to these
questions, and agreed, for example, that the cause of a sealed container exploding
its contents out is the internal pressure caused by the increase in the heat of the
fluid inside the container, that this explosion is unintentional, and occurs in a
violent manner.
These brief responses provide a rough, nonlinguistic profile of people’s
understanding of a particular source domain concept (i.e. heated fluid in the bodily
container). A significant part of this knowledge comes from people’s own
© 2006 The Author
Journal compilation © 2006 Blackwell Publishing Ltd
 Metaphor Interpretation as Embodied Simulation 439

understanding of their bodies. For instance, with the example of heated fluid in a
container, children come to understand this idea early in infancy as they feel things
going in and out of their bodies, and experience their bodies going in and out of
containers (e.g. cribs, beds, rooms, cars, etc). Children experience their bodily
fluids (i.e. sweat, blood, urine) as being differentially heated and under pressure
depending on various, often emotional, circumstances. Thus, people do not need
to understand the physics of how heated fluid behaves in sealed containers out in
the external world to have some intuitive, and embodied, understanding of why
anger is sometimes metaphorically characterized as heated fluid in a container.
Are people’s understandings of anger idioms (e.g. ‘blow your stack’) partly
structured by their folk concept for heated fluid in the bodily container as described
above? To test this possibility, people in a second study were presented with stories
ending in either an idiomatic phrase or literal paraphrase, as shown in the following
example:

Sally was preparing for a big dinner party.

She had to do a great deal of cooking.
Her husband was supposed to help, but was very late getting home from
work.
When her husband strolled in 10 minutes before the party whistling and
smiling,
Sally blew her stack
or
Sally got very angry.

After reading each story, and either the final metaphorical or literal expression,
participants gave ratings of agreement for three statements that were created
directly from the image-schematic structure of the source domain as determined
by the results of the first study. Thus, for the above story, people saw three
statements in regard to the causes, intentionality, and manner by which Sally
exhibited her anger:

(a) Sally got very angry because she was under a great deal of pressure
(causation);
(b) Sally got very angry without intending to do so (intentionality);
(c) Sally got very angry in a forceful manner (manner).

People gave their ratings of agreement to each of these statements, on a 7-point

scale (with 7 meaning ‘strong agreement’ and 1 meaning ‘little agreement’). Given
that people read the same story in both the metaphoric and literal conditions, any
difference in their ratings reflects their conceptual understanding of last phrase
read. The prediction was that people would give higher ratings of agreement to
the three statements having read the idioms than the literal paraphrases, precisely
because of their tacit understandings that these idioms were motivated by embodied
© 2006 The Author
Journal compilation © 2006 Blackwell Publishing Ltd
440 R. W. Gibbs

conceptual metaphors. Not surprisingly, when people understand anger idioms,

such as ‘blow your stack’, ‘flip your lid’, or ‘hit the ceiling’, they gave significantly
higher agreement ratings to the causation, intentionality, and manner statements
when they read idioms than when they saw literal paraphrases. This suggests that
when people read idioms like ‘blow your stack’ they inferred that the cause of
anger is internal pressure, that the expression of anger is unintentional, and is done
is an abrupt violent manner. People did not draw these same inferences about
causation, intentionality, and manner when comprehending literal paraphrases of
idioms, such as ‘get very angry’, primarily because people can ‘get very angry’ in
many ways without experiencing heat. Two control studies examined, and
rejected, alternative hypotheses that these rating differences were due to the literal
paraphrases per se, or to the different verbs employed in the idioms and the literal
phrases.
In general, these findings indicate how the metaphorical mappings between
embodied source domains and target domains motivate the specific figurative
meanings of many idioms, and preserves the structural characteristics, or cognitive
topology, of the source domains. More importantly, these data also showed that
people have specific metaphorical conceptions of many abstract ideas (e.g.
emotions) that are shaped, to some extent, by recurring bodily experiences (e.g.
their own bodies as containers), in addition to their encounters with real-world
scenes (e.g. observing the behavior of closed pots containing boiling water).
Other psycholinguistic research demonstrates that people’s specific intuitions
about their bodily experiences of hunger provide the grounding for the acceptable
use and understanding of metaphorical expressions in two languages (American
English and Brazilian Portuguese) referring to different kinds of human desires
(Gibbs, Lima, and Francuzo, 2004). For example, American English speakers
frequently talk of abstract desires in terms of hunger, such as in ‘He hungers for
recognition’ or ‘He hungers for adventure’. Asserting this metaphorical relationship
is not just a conventional or arbitrary way of speaking about desire, because there
appear to be rich, systematic correspondences between feeling hunger and feeling
different aspects of desire, including lust/sex, concrete objects, and abstract
idea/events. This idea of DESIRE IS HUNGER reflects a pervasive ‘primary
metaphor’ (Grady, 1999) given the high, but not at all perfect, degree of positive
correlation in people’s bodily experiences of hunger and desire. Gibbs et al. (2004)
showed how knowing something about people’s embodied concepts of hunger
allows one to empirically predict which aspects of desire will, and will not, be
thought of, and talked about, in specific metaphorical terms across two different
languages and cultural communities. Once again, people use their bodily
experiences/actions as primary sources for metaphorical concepts and their use and
understanding of metaphoric language referring to these ideas.
None of the studies described above necessarily indicate that people used their
‘in-the-moment’ felt sense of their bodies when thinking about and understanding
the metaphorical language referring to abstract concepts. Yet other work shows
that people at least appear to access embodied conceptual metaphors in some form
© 2006 The Author
Journal compilation © 2006 Blackwell Publishing Ltd
 Metaphor Interpretation as Embodied Simulation 441

during immediate metaphor processing (Gibbs, Bogdonovich, Sykes and Barr,

1997; Pfaff, Gibbs, and Johnson, 1997). As noted above, the linguistic and
psycholinguistic research in support of conceptual metaphors is widely debated
with several critics voicing concern about the methodological problems in
discerning conceptual metaphors from systematic patterns of linguistic expressions
and even whether conventional expressions such as ‘Our marriage is at the cross
roads’ and ‘John blew his stack’ convey metaphorical meaning (Glucksberg, 2001;
Haser, 2005; Kennedy and Veraerke, 1993; Keysar, Shen, Horton and Glucksberg,
2000). It is not my intention here to address these important issues other than to
properly acknowledge them as questions for future discussion and research (see
Deignan, 2005; Gibbs, 2006a for some critical responses to these arguments against
conceptual metaphors). I do, however, want to extend the idea that embodied
conceptual metaphors are critical to understanding many kinds of metaphoric
language by arguing for the importance of embodied simulation in metaphor
interpretation.

3. Simulations and Cognitive Processing

Cognitive linguists have mostly characterized the activation of conceptual metaphor

during metaphor understanding as a purely cognitive process. Thus, understanding
the conventional phrase ‘Our relationship has hit a dead-end street’ is partly
accomplished through the activation of the conceptual metaphor LIFE IS A
JOURNEY in long-term memory. This enduring chunk of metaphorical
knowledge has a source domain (e.g. JOURNEY) that is grounded in the pervasive
bodily experience, or image-schema, of SOURCE-PATH-GOAL. But the entire
process of accessing a specific conceptual metaphor during verbal metaphor
understanding is mostly viewed as activating abstract, schematic, disembodied
knowledge that is not tied to ongoing bodily action.
My claim is that understanding what many words and phrases mean requires
that listeners engage in an experiential/embodied simulation of the described
situation. Consider, for example two different headlines to news articles posted on
the internet: ‘Parrot prodigy may grasp the concept of zero’ and ‘Journalists who
grasp the concept of courage’ (http://www.nationalgeographic.com/news/2005/
07/07/05 and http://www.ohio.com/old/ohio/2005/11/13/news/editorial
respectively). Both headlines describe a physical action of grasping in the context
of abstract entities that are impossible to physically touch or control (i.e. the
concepts of ‘zero’ and ‘courage’). At first glance, it seems odd to associate the
notion of grasping with nonmaterial ideas. But cognitive linguistic research has,
again, shown how people ordinarily conceive of abstract concepts in physical terms
and can apply various embodied actions to these objects/concepts as a result. For
example, concepts are not physical objects that can be touched, held on to, juggled,
dropped, and so on. When hearing ‘grasp the concept’ listeners engage in, or
imagine engaging in, a relevant body action, such as grasping, that facilitates
© 2006 The Author
Journal compilation © 2006 Blackwell Publishing Ltd
442 R. W. Gibbs

metaphorical construal of the abstract notion of ‘concept’ as a kind of physical

entity, such that concepts can indeed be things that are grasped, held on to,
dropped, misplaced, chewed on, and so on. Conceiving of abstract entities as
physical objects enables people to perform mental actions on these objects as if
they possessed the properties of real-world, concrete, physical entities. In this way,
partial reenactment of sensorimotor processes related to ‘grasping’ underlies
conceptual knowledge, reasoning, and linguistic understanding.
This perspective on metaphor understanding, more specifically, as a process of
embodied simulation is contrary to the typical view of metaphor understanding as
based on activating abstract knowledge that may have arisen from bodily activity, but
is currently represented in an amodal, and disembodied manner. At the same time,
this hypothesis departs radically from the traditional belief that action phrases such as
‘grasp the concept’ have meanings that were once transparently metaphoric, but over
time have lost their original metaphoricity and now exist as clichés or dead metaphors
(see Gibbs, 1994 for discussions of this view), or that the physical meaning of ‘grasp’,
for instance, must be inhibited in the context of ‘grasp a concept’. Under this
traditional view, people know that the phrase ‘grasp the concept’ refers to understanding
something, but are unaware as to why this phrase has the particular meaning it does
(i.e. why it is that concepts are physical entities that can be ‘grasped’).
The main point of constructing an embodied simulation is to create a sense of
what it must be like for others, such as speakers or writers, to have the specific
thoughts they do during communication. Many metaphorical thoughts are
embodied in the sense that these arise from bodily experiences that people
imaginatively engage in as they speak. Most simulation theorists assume that the
pretend state created during simulation arises from a deliberate and voluntary act
(Gordon, 1986; Harris, 1989; Goldman, 2001). But simulation processes are
different from engaging in pretense (Currie and Ravenscroft, 2002), and may be
automatic, unconscious, and pre-reflexive (Gallese, 2003). Thus, one pretends to
do something (e.g. talking on a telephone) by performing some other, somewhat
analogous, action (e.g. holding a banana up to one’s ear). On the other hand,
imaginative simulations are mental actions where one is not doing one thing to
stand for another, but where one mentally engages in actions similar to those
overtly referred to. For instance, when I imagine what it feels like to kick a ball, I
do not engage in some other action, such as kicking a rock to do so. Instead, I
mentally construct a scenario of my own body kicking a ball. This simulation is
not abstract, in the way, for example, that a computer simulation of a hurricane
mimics abstract elements of how a hurricane moves. Embodied simulations have a
full-bodied feel to them, in the way that a person may experience actual sensations
of movement when flying an aircraft simulator. People may not necessarily be
aware of these sensations, even if research on ideomotor actions demonstrates that
people often unconsciously move in similar patterns to others around them (Knuf,
Aschersleben and Prinz, 2001). Simulations are imaginative acts that are intimately
involved with subpersonal processes, (Currie and Ravenscroft, 2002) and in most
cases are performed automatically without significant conscious reflection. Of
© 2006 The Author
Journal compilation © 2006 Blackwell Publishing Ltd
 Metaphor Interpretation as Embodied Simulation 443

course, there are times when we will deliberately imagine engaging in some bodily
recreation of something. Yet there is a variety of research showing that ordinary
perception of action is equivalent to internally simulating it.
For example, early studies showed shared mechanisms for action perception and
action control in monkeys. ‘Mirror neurons’ in monkey ventral premotor cortex
are active both when a monkey observes a specific action, such as someone grasping
a food item, and when the monkey performs the same kind of action (Gallese,
2000). Neurons in monkey premotor cortex discharge both when the animal
performs a specific action and when it hears the corresponding action-related
sound (Kohler, Keysers, Umilta, Fogassi, Gallese and Rizzolatti, 2002). Observations
like this have been extensively reported in human studies showing that there are
shared motor representations for action, observation of another person’s actions,
and imitation and mental simulation of action (Decety and Grezes, 1999).
Simulation is generally important in mental imagery, action understanding,
imitation, and empathy (Gallese, 2003), as well as creating metaphorical concepts
(Gallese and Lakoff, 2004). More recently, it has been shown that the mirror
neurons system is directly involved in the perception of communicative action
(Buccino, Lui, Canessa, Patteri and Lagravinese, 2004), in imitation (Koski,
Iacoboni, Dubeau, Woods and Mazziotta, 2003) and in basic forms of mind
reading (Jellema, Baker, Wicker and Perrett, 2000; Umilta, Kohler, Gallese, Fogassi
and Fadiga, 2001).
These neuroscientific studies support the idea that perceptual and motor systems
are specifically activated during people’s understandings of others’ actions and the
communicative thoughts that motivate these acts. Other research suggests that
perceptual and motor systems are activated during immediate language processing.
Thus, areas of motor and pre-motor cortex associated with specific body parts are
activated when people hear language referring to those body parts. Listening to
different verbs associated with different effectors (i.e. mouth/’chew’, leg/’kick’,
hand/’grab’) leads to different firing rates in different regions of motor cortex (i.e.
areas responsible for appropriate mouth/leg/hand motions exhibit greater
activation) (Hauk, Johnsrade and Pulvermuller, 2004; Pulvermuller, Haerle and
Hummel, 2001).
Psycholinguistic studies also demonstrate the automatic recruitment of perceptual
and motor systems in immediate language understanding. For instance, reading
sentences with visual semantic components can selectively interfere with visual
processing. Thus, participants took longer to perform a visual categorization task
in the upper part of their visual field when they heard sentences depicting upward
motion such as ‘The ant climbed’ (Richardson, Spivey, McRae and Barsalou,
2003). When people perform physical actions such as forming a fist or moving a
lever toward the body, they were slower to verify as meaningful sentences that
described unrelated actions, such as ‘aim a dart’ (Klatzky, Pelligrino, McCloskey
and Doherty, 1989), and ‘close the drawer’ (Glenberg and Kaschak, 2002). Finally,
psycholinguistic studies show that people understand fictive motion sentences,
which express metaphorical meaning, such as ‘The road runs along the coast’, in
© 2006 The Author
Journal compilation © 2006 Blackwell Publishing Ltd
444 R. W. Gibbs

terms of implicit, imaginary sensations of movement (Matlock, 2004). Thus,

people interpret the meanings of fictive motion statements by ‘replaying’ the
movement (i.e. physical running), reconstructing a mental enactment of movement
implicit in the sentence. People are not aware of these simulations, and so fictive
motion processing is not dependent on deliberate thought about motion. It is
possible that the imaginary sensations of movement that accompany processing of
both metaphorical and nonmetaphorical statements are simply the products of
understanding and not part of the comprehension process itself. Although the
above studies do not explore this idea in detail, the priming results observed in
these studies, and the findings from the mirror neuron studies, suggest that
simulation processes are important to the immediate construal of meaning and not
just an afterthought to understanding.
In general, these psycholinguistic studies provide additional support for the
broad claim that language use is closely tied to embodied imagination. There is
also an emerging body of research from computational modeling, known as
‘simulation semantics’, that gives a primary role to embodied simulations in
drawing appropriate inferences from various metaphorical and non-metaphorical
language (Bergen, 2005; Feldman and Narayanan, 2004). The remainder of this
paper describes three research projects, some of which are still in progress, which
provide additional behavioral evidence on embodied simulations in metaphoric
language interpretation. My concern in the following section is to show how
people’s felt experiences of their bodies in action underlies their ability to bodily
imagine many kinds of metaphorical actions referred to in much metaphoric
language.

4. Simulation and Metaphor Understanding

4.1 Imagining Metaphorical Actions

It is easy to form mental images of people engaging in concrete actions, such as
chewing the gum, swallowing a pill, coughing up blood, and breaking off a branch.
But is it possible to form coherent mental images for physically impossible actions,
such as chewing on an idea, swallowing one’s pride, coughing up a secret, or
breaking off the relationship? Previous studies have demonstrated the utility of
examining people’s mental images for figurative language as a window into the
type of conceptual knowledge needed to makes sense of metaphorical phrases,
such as idioms (e.g. ‘spill the beans’, ‘let the cat out of the bag’, ‘blow the lid off’)
(Gibbs and O’Brien, 1990) and proverbs (e.g. ‘A rolling stone gathers no moss’ and
‘Let sleeping dogs lie’) (Gibbs, Strom and Spivey-Knowlton, 1997). These studies
demonstrate that people are limited in the kinds of images they create for idioms
and proverbs because of very specific embodied knowledge that helps structure
their metaphorical understanding of various concepts. For example, people’s
images for anger idioms (e.g. ‘blow your stack’ and ‘flip your lid’) are based on folk
conceptions of certain physical events. That is, people use their embodied
© 2006 The Author
Journal compilation © 2006 Blackwell Publishing Ltd
 Metaphor Interpretation as Embodied Simulation 445

knowledge about the behavior of heated fluid in containers (e.g. their bodies as
containers and bodily fluids within them) and map this knowledge onto the target
domain of anger to help them conceptualize in more concrete terms the concept
of anger (i.e. ANGER IS HEATED FLUID IN A CONTAINER). Various
specific meanings result from these general metaphorical mappings, ones that
provide specific insight into people’s consistent responses about the causes,
intentionality, manner, and consequences of the activities described by stacks
blowing, lids flipping, ceilings being hit and so on. It appears, then, that the
embodied metaphorical ways in which people partially conceptualize experiences
actually provide part of the motivation for why speakers have consistent mental
images and specific knowledge about these images for idioms and proverbs with
similar figurative meanings.
One new set of studies investigated whether people can form mental images for
metaphorical phrases that refer to actions literally impossible to perform, such as
‘chewing on the idea’, ‘stretch for understanding’, and ‘cough up a secret’ (Gibbs,
Gould and Andric, in press). These studies attempted to see if mental imagery for
metaphor provides evidence of embodied simulation. Our hypothesis was that
people’s mental images should provide evidence of the embodied character of
metaphor understanding, even when the actions referred to are impossible to
perform. People’s embodied metaphorical conceptualizations of abstract ideas
should enable them to imagine the ways that abstract entities may be acted upon.
Unlike imagining nonmetaphoriocal action statements (e.g. ‘chew on the gum’),
where people’s images should focus on the procedural characteristics of the
concrete actions (i.e. moving their mouths as they chew the gum), people’s mental
images for metaphorical phrases should show an analogical understanding of how
abstract domains, such as ideas or concepts, can be actively structured in terms of
embodied source domains (i.e. chewing on something to get more out of it).
Furthermore, having people watch, imitate, or imagine engaging in relevant
embodied actions (e.g. chewing or grasping) may enhance the degree to which
they conceptualize metaphorical actions through embodied simulations, especially
when these relevant embodied actions are performed or imagined before they hear
a metaphoric phrase (e.g. ‘grasp the concept’).
The results of a first study showed that when participants were presented specific
phrases that were either metaphorical (e.g. ‘grasp the concept’) or nonmetaphorical
(e.g. ‘grasp the branch’), and given ten seconds to form a mental image of that
action, they rated it easier to form and easier to feel mental images for
nonmetaphorical phrases than metaphorical ones. Not surprisingly people also
rated that it was more difficult to imagine actually performing the actions in
metaphorical phrases than was the case for the nonmetaphorical expressions.
The remaining questions more directly probed the contents of people’s mental
images for the metaphorical and nonmetaphorical phrases. When asked ‘What is
particularly noticeable in your image?’ people gave a wide variety of responses to
this question, and these could be roughly divided into two groups. The first set of
answers made some specific reference to the participants actually participating in
© 2006 The Author
Journal compilation © 2006 Blackwell Publishing Ltd
446 R. W. Gibbs

the action. For example, when one participant was given the metaphorical phrase
‘chew on the idea’, she noted ‘My jaw goes up and down as I chew’. People
actually gave far more of these specific references to participating in the action
responses for the nonmetaphors (63%) than to the metaphors (29%). This result
makes sense given that people found it easier to image the specific action performed
in the nonmetaphorical phrases.
The other set of responses provided a conceptualized description of the action.
For instance, when a person was presented with the metaphor ‘stretch for
understanding’, he said that the most noticeable thing in his image was ‘there is
much stretching going on both in terms of the ideas being stretched out to see if
they are true and me stretching to better see or examine the idea’. This response
provides an excellent example of how embodied metaphor constrains the kinds of
mental images people construct when hearing metaphorical action statements. The
participant essentially noted that IDEAS ARE OBJECTS which can be physically
inspected, by stretching them out to more effectively examine them, and that
UNDERSTANDING IS GRASPING enables the person to extend his or her
body to better control the object, and thus better understand it. Overall, people
gave significantly more conceptualized description of the action responses when
they were presented with the metaphors (71%) than the nonmetaphors (37%).
This too makes sense given that the nonmetaphors can be understood simply in
terms of the physical actions they depict without any mental mapping of knowledge
from a different domain of experience. The metaphors, on the other hand, are
understood in terms of mappings from embodied source domains onto abstract
target domains, which influence what participants find most notable in their mental
images for these figurative phrases.
The final question asked participants was ‘Why is this concept (e.g. idea)
sometimes associated with this action (e.g. chewing)?’. Once again, participants
gave a wide range of answers in response to this question, and these were roughly
classified into two groups. The first set of responses focused on providing a concrete
explanation of the relevant process or action. For instance, when one participant
heard the nonmetaphorical phrase ‘chew on the gum’. she responded with ‘That
is what you do with gum—chew on it’. Another participant was presented with
the metaphorical phrase ‘put your finger on the truth’ and responded ‘because that
is what you do in explaining the truth’. Both these responses concentrate on the
conventional relationship between some object or concept and some action or
process. Yet none of these responses provide in-depth insights into the underlying
relationship between a concept and an action. The data revealed that participants
gave significantly more concrete explanation of the process answers to the
nonmetaphors (59%) than to the metaphors (19%).
The other group of answers specifically provided an analogous, conceptual
explanation as to why some concept was sometimes associated with some action
or process. For instance, for the metaphorical phrase ‘chew on the idea’, one
person said ‘Chewing is related to a slow methodological activity and it could be
related to turning something over in your mind to better understand it’. Overall,
© 2006 The Author
Journal compilation © 2006 Blackwell Publishing Ltd
 Metaphor Interpretation as Embodied Simulation 447

people gave analogous, conceptual explanations significantly more often to the

metaphors (77%) than to the nonmetaphors (36%). This result is as expected,
assuming that people’s mental images for metaphorical action phrases are constrained
by their embodied, metaphorical understandings of the target domains referred to
in these expressions (e.g. ideas, concepts, feelings). Nonmetaphorical phrases are
not understood via embodied metaphorical knowledge in quite the way that
metaphorical phrases are, and people are therefore far less likely to focus on these
underlying conceptual, embodied motivations in constructing mental images for
realistic physical actions, such as chewing the gum.
A second study more directly explored the possibility that people imagine
metaphorical actions by engaging in embodied simulations of the actions referred
to in the metaphorical statements. Participants once again heard different
metaphorical and nonmetaphorical expressions, formed mental images for these
phrases, and then answered a series of questions about their images. In Experiment
2, however, people also participated in one of three enactment conditions in which
they first did one of three things: (a) watched the experimenter make a bodily
action relevant to the main verb in each statement (e.g. making a stretching motion
before forming a mental image for the phrase ‘stretch for understanding’), (b)
watched the experimenter make a relevant bodily action, which they then imitated,
before being given ten seconds to form their mental image for a phrase, or (c)
watched the experimenter make a relevant bodily action, then imagined themselves
doing the same action, before forming a mental image for the phrase. These three
experimental treatments were referred to as the watching, imitating, and imagining
conditions, respectively.
We included these different enactment conditions to see whether people’s
understanding and imagining of metaphorical phrases would be enhanced relative
to that noted in Experiment 1 where people formed mental images without
explicitly engaging in some sort of relevant bodily engagement. The embodied
simulation view of metaphor understanding suggests that it is perfectly plausible
to interpret ‘stretch’ in a conventional way, referring to a specific bodily action,
precisely because the notion of ‘understanding’ is quickly recognized in an
embodied metaphorical way as something that can be physically touched and
grasped (e.g. UNDERSTANDING IS ACQUIRING POSSESSION OF AN
OBJECT). Thus, conventional aspects of word meaning are recruited as part
of the moment-to-moment process of partially reenacting the action referred
to by the statement ‘stretch for understanding’. In a similar manner, people’s
understanding of ‘stretch for understanding’ should result in a mental image that
includes something of the products of this embodied simulation. Having people
actually stretch, watch someone else stretch, or imagine themselves stretching,
should, therefore, not interfere with people’s understanding of the metaphorical
meanings of the metaphors, or their explanations of why these phrases mean
what they do.
One possibility is that engaging in different forms of bodily action may lead to
greater awareness of the embodied conceptual processes used to comprehend
© 2006 The Author
Journal compilation © 2006 Blackwell Publishing Ltd
448 R. W. Gibbs

phrase like ‘stretch for understanding’, compared to when people formed mental
images for these phrases without bodily engagement (e.g. Experiment 1). The
main difference between Experiments 1 and 2 was that people engaging in various
forms of bodily action in Experiment 2 viewed the metaphorical phrases in some
cases as easier to image, in the watch condition, and to feel, in the imitate condition,
than they did the nonemtaphorical phrases. It appears that some aspects of bodily
engagement enhance, to some degree, people’s abilities to form more robust
images for the metaphorical action phrases.
Analysis of the responses to the conceptual explanations for the metaphors,
collapsed across all three enactment conditions, showed that 78% of these referred
to additional bodily actions and consequences of these actions related to the main
verb in each metaphorical phrase. This proportion was higher than that obtained
in Experiment 1 (48%). For example, when one participant was given the phrase
‘put your finger on the truth’ (in the imagine condition), she replied ‘I guess being
able to touch the truth is an important thing, being able to relate to it, being able
to actually see that it is a physical thing and can be examined’. This evidence
reflects the product of the embodied simulation the participant constructed that
made this impossible action plausible and meaningful.
There are several important conclusions to be draw from these findings. First,
people’s mental images for metaphorical actions provide an interesting type of
evidence about their embodied understandings of what these verbal phrases
metaphorically mean. People’s responses to questions about their images for
metaphorical action phrases may reflect part of the embodied simulations they
engage in when interpreting these phrases. Second, people can form mental images,
and have distinct intuitions about them, for metaphoric phrases referring to
physically impossible actions such as chewing on ideas, stretching for understanding,
tearing apart the argument, etc. These mental images were not static, but exhibited
significant unfolding and transformation as if people were viewing or experiencing
them in real time. Finally, the images here were not purely visual, as participants
primarily noted other sensory and kinesthetic characteristics of the actions referred
to in the metaphorical phrases. Metaphorical actions, despite their physical
impossibility, have a ‘feel’ to them, because of people’s tacit, and in these
experimental tasks, explicit understanding of their full-bodied meanings.
We do not claim that people necessarily engage in these embodied, imaginative
processes every time they hear these familiar metaphorical phrases. After all, the
highly conventional nature of many of the metaphorical phrases studied here
suggests that these are easily understood, and people may not always form explicit
mental images of metaphorical language, or any other type of language for that
matter, during on-line processing of these expressions in ordinary discourse. The
extent to which people ordinarily engage in imagistic processes during online
comprehension is an important topic for future research. But the present studies
show that asking people to slow down and form explicit mental images for
metaphoric and nonmetaphorical language can reveal significant differences in
people’s intuitions about why these phrases have the specific meanings they do,
© 2006 The Author
Journal compilation © 2006 Blackwell Publishing Ltd
 Metaphor Interpretation as Embodied Simulation 449

including the important role of embodied metaphorical thought in constructing

simulations of different metaphorical actions. Having people engage in different
kinds of bodily action does not interfere with the creation of metaphorical images,
and in some cases seems to highlight the imaginative reasons for why impossible
metaphorical action phrases are meaningful and plausible.

4.2 Bodily Movement and Metaphor Comprehension

A different line of research investigated the possible influence of bodily action on
the speed with which people process simple metaphoric phrases, as ‘stamp out a
feeling’, ‘push an issue’, ‘sniff out the truth’ and ‘cough up a secret’, each of which
denote physical actions upon abstract items. Wilson and Gibbs (2005) hypothesized
that if abstract concepts are indeed understood as items that can be acted upon by
the body, then performing a related action should facilitate sensibility judgments for
a figurative phrase that mentions this action. For example, if participants first move
their arms and hands as if to grasp something, and then read ‘grasp the concept’,
they should verify that this phrase is meaningful faster than when they first
performed an unrelated body action. Our hypothesis was that engaging in body
movements associated with these phrases should enhance the online simulations
that people create to form a metaphorical understanding of abstract notions, such as
‘concept’, even if ‘concepts’ are not things that people can physically grasp.
Participants in a first study initially learned to perform various specific bodily
actions (e.g. throw, stamp, push. swallow, cough, grasp) given different nonlinguistic
cues. Following this, participants were individually seated in front of a computer
screen. The experiment consisted of a series of trials where an icon flashed on the
screen, prompting the participant to perform the appropriate bodily action. After
doing this, a string of words appeared on the screen and participants had to judge
as quickly as possible whether that word string was ‘sensible’.
Analysis of the speeded sensibility judgments showed that participants responded
more quickly to the metaphorical phrases that matched the preceding action (e.g.
the motor action grasp was followed by ‘grasp the concept’), than to the phrases
that did not match the earlier movement (e.g. the motor action kick was followed
by ‘grasp the concept’). People were also faster in responding to the metaphor
phrases having performed a relevant body movement than when they did not
move at all. In short, performing an action facilitates understanding of a figurative
phrase containing that action word. A second study showed that same pattern of
bodily priming effects when participants were asked to imagine performing the
actions before they made their speeded responses to word strings. This result
reveals that real movement is not required to facilitate metaphor comprehension,
only that people mentally simulate such action.
One possibility for these studies is that having people perform specific actions or
imagine doing so, such as coughing, prompted them to think of the word ‘cough’
that in turn primed their recognition of the phrase containing that word (e.g.
‘cough up a secret’). But a separate control study showed that there was no
© 2006 The Author
Journal compilation © 2006 Blackwell Publishing Ltd
450 R. W. Gibbs

correlation between the frequency with which people could come up with the
identical word for some action after they performed the action (i.e. mimicking the
action performed by the actor on the videotape) and the reading time results for
phrases containing those words. This finding suggests that the priming effects
noted in these two reading time studies were not due to simple lexical associations
created from doing or imagining the actions and seeing specific words in the
phrases.
Most generally, people do not understand the nonliteral meanings of these
figurative phrases merely as a matter of convention. Instead, people actually
understand ‘toss out a plan’, for instance, in terms of physically tossing something
(i.e. the plan is viewed as a physical object). In this way, processing metaphoric
meaning is not just a purely cognitive act, but involves some imaginative
understanding of the body’s role in structuring abstract concepts. People may create
embodied simulations of speakers’ messages that involve moment-by-moment ‘what
must it be like’ processes that make use of ongoing tactile-kinesthetic experiences.
These simulation processes operate even when people encounter language that is
abstract, or refers to actions that are physically impossible to perform.
At the same time, there are important communicative reasons for why people
actually move their bodies in particular ways when speaking that may be directly
related to attempts to facilitate people’s comprehension of metaphor. Studies of
people, both experts and novices, speaking about abstract topics (e.g. mathematics,
physics) show that appropriate physical gestures are often employed to articulate
something about the topic at hand (McNeil, 1992; Roth and Lawless, 2002).
Thus, mathematicians exhibit gestural images for the concept of limits, both direct
and inverse (e.g. hand moving a straight line in front of the body for direct limits,
and hand looping downward and back up for inverse limits) (McNeil, 1992).
These gestures embody (‘give a body to’) abstract, metaphorical ideas, and
sometimes precede the language spoken to enhance listeners’ understandings of
speakers’ complex, abstract communicative intentions. Similarly, people may reach
out their hand and grasp the air before, saying ‘I finally grasped that concept’ to
facilitate addressees’ comprehension of the metaphorical idea of grasping a concept.
It would be interesting to explore whether the timing of such gestures influenced
the speed with which people comprehended metaphorical action phrases. Thus,
making relevant body movements before reading metaphorical phrases may
facilitate their comprehension. Making the gestures during reading, on the other
hand, might interfere with processing because the gesture and the simulation
process engaged to comprehend the metaphorical phrase would compete for neural
resources in motor cortex.

4.3 Embodied Simulations in Understanding Metaphorical Narratives

A final set of studies examined people’s embodied interpretation of simple
metaphorical narratives (Gibbs, 2006b). Consider the following two brief narratives
about the development of two different romantic relationships.
© 2006 The Author
Journal compilation © 2006 Blackwell Publishing Ltd
 Metaphor Interpretation as Embodied Simulation 451

Story A
Imagine that you are a single person. A friend sets you up on a blind date. You
really like this person and start dating a lot. Your relationship was moving along in
a good direction. But then it got even better. The relationship felt like it was the
best you ever had. This continues to this day. No matter what happens, the two
of you are happy together.

Story B
Imagine that you are a single person. A friend sets you up on a blind date. You
really like this person and start dating a lot. Your relationship was moving along in
a good direction. But then you encountered some difficulties. The relationship did
not feel the same as before. This lasted for some time. No matter how hard you
two tried, the two of you were not getting along.

Story A and B differ in that A describes a successful relationship, while B

describes a relationship that appears to be in trouble. Both stories are similar,
however, in conceiving of the relationships as entities that can move along some
sort of path, as indicated in the fourth line ‘Your relationship was moving along in
the good direction’. Although no other part of the two stories explicitly refers to
journeys, the two stories provide different impressions of the ‘relationship journey’.
Thus, Story A suggests a smooth, uninterrupted journey that is still progressing,
while Story B implies a more difficult, perhaps interrupted, journey that may no
longer be progressing.
It is reasonable to talk about abstract entities like romantic relationships in terms
of bodily actions like ‘moving along in a good direction’ precisely because many
abstract concepts are partly understood in terms of enduring embodied metaphors
such as RELATIONSHIPS ARE JOURNEYS. This metaphorical mapping gives
rise to a range of meaning correspondences, such as that relationships can commence
in varying ways at different speeds, encounter obstacles along some path, face
turning points, can head toward specific destinations, and be judged as successful
or not by the progress made along the journey.
The goal of one new set of studies was to provide evidence, using a new
methodology, showing that people’s interpretations of simple narratives partly rely
on their embodied understandings of the metaphors involved (Gibbs, 2006b). My
hypothesis, once more, was that people infer the detailed meanings of simple
narratives involving conceptual metaphors by imagining their participating in the
metaphorical actions explicitly mentioned in these stories. For example, when
hearing ‘moving along in a good direction’, listeners imagine engaging in a body
action, such as traveling along some path, which facilitates their metaphoric
understanding of the abstract, and physically impossible, idea that romantic
relationships can move along a path toward some goal. If people imaginatively
simulate themselves in the journey, then listening to these different renditions of
the RELATIONSHIPS ARE JOURNEYS conceptual metaphor should have
different real world embodied effects. To assess this idea, people listened to a story,
© 2006 The Author
Journal compilation © 2006 Blackwell Publishing Ltd
452 R. W. Gibbs

were blindfolded, and then walked along a path toward an object while they
thought about the story. People should walk differently when hearing successful
and unsuccessful metaphor stories, while these effects should be greatly attenuated
after hearing nonmetaphorical narratives that did not suggest a conceptualization
of the relationship as a kind of physical journey.
Before conducting the main walking experiment, a first study examined the
meanings people inferred from reading the metaphorical and nonmetaphorical
stories. University students read one of two pairs of stories and then answer a series
of questions that were designed to tap into their intuitions about what they read.
One pair of narratives was the two stories presented above containing a metaphorical
statement ‘Your relationship was moving along in a good direction’ in which one
story ultimately described a successful relationship and one that was unsuccessful.
The other pair of stories was identical to the first pair, but had the key metaphorical
statement ‘Your relationship was moving along in a good direction’ replaced by
the nonmetaphorical expression ‘Your relationship was very important to you’.
After reading one pair of stories, participants answered a series of questions about
the spatial characteristics of each relationship. My expectation was that people
would see the two metaphorical stories to be quite different, precisely because of
their embodied understanding of the ‘Your relationship was moving along in a
good direction’ statement. However, they would find the pair of nonmetaphorical
statements to be less different and less likely to draw the same kinds of inferences
as when they read different metaphorical stories.
In fact, participants’ choices revealed that they found the successful metaphorical
relationships to be progressing further, but not at the beginning, moving along in
a straighter line, and the story participants to be heading more in the same direction
than was the case for the unsuccessful metaphorical story. It is important to note
that there is nothing in the metaphorical stories that directly assert anything about
the distance, speed, extent, and direction of the relationship ‘journeys’ traveled. All
of these inferences were drawn on the basis of people metaphorical understandings
of the stories as referring to RELATIONSHIPS ARE JOURNEYS, as suggested
by the ‘Your relationship was moving along in a good direction’ statement. The
data clearly suggest that the couple in the successful story had progressed further
overall, were progressing faster at the present time, were moving more along a
straight path, and were headed in the same direction, compared to the couple
depicted in the unsuccessful story.
Participants’ responses to the same questions for the nonmetaphorical stories
were different than seen for the metaphorical narratives. People significantly chose
successful stories more often in response to the questions about the ‘progressing
further’ and ‘heading in the same direction’ questions. These findings generally
suggest that people can draw some metaphorical inferences about the distance,
speed, and direction of romantic relationships from nonmetaphorical stories. Yet,
more importantly, people draw far more specific, and consistent metaphorical
inferences about the nature of relationship ‘journeys’ when reading metaphorical
as opposed to nonmetaphorical narratives.
© 2006 The Author
Journal compilation © 2006 Blackwell Publishing Ltd
 Metaphor Interpretation as Embodied Simulation 453

The main experiment more closely examined whether people’s specific

embodied understandings of the metaphorical and nonmetaphorical narratives may
have been produced through their imaginative simulations of the different
relationship ‘journeys’. This possibility was tested in a novel manner by having
participants physically walk toward an object, 40’ away, after hearing either a
successful or unsuccessful story in either the metaphorical or nonmetaphorical
condition. Based on the findings from Experiment 1, my hypothesis was that
participants’ understanding of the stories involved an imagined reenactment of the
metaphorical actions (e.g. moving along in a good direction) which should lead
them to walk longer and farther having heard the successful relationship story than
the unsuccessful story. This same pattern of walking should be greatly reduced
when participants heard the two versions of the nonmetaphorical narratives. Once
they stopped, an experimenter nearby asked the participants to rate on a 7-point
scale how they felt at the moment, and to what extent they were actually thinking
about the story as they walked or imagined walking toward the target. The
experimenters then measured how close the participants were from the yellow
ball, and how far each participant wandered from the straight line from the starting
point to the yellow ball.
Furthermore, another experimental condition asked participants to simply
imagine walking to the object after hearing one of the stories. For the imagined
condition, participants were blindfolded, but instructed to only imagine walking
out to the yellow ball as they thought about the story, and to press a stopwatch as
soon as they imagined arriving at the ball. Once a participant had stopped the
watch, the experimenter asked the same questions noted above. A similar pattern
of results should be found in the imagine condition as in the walk condition, given
earlier data showing strong equivalence between imagined and real action (Grezes
and Decety, 2001).
Participants in the walk condition did not always walk in a direct straight line
from the starting point toward the target. The experimenters, therefore, measured
both the distance from where the participant stopped to the straight line linking
the starting point and the target, as well as the distance from that point along the
vertical line to the target. Analysis of the horizontal distances from the starting
point-to-target line showed no differences among the different experimental
conditions, and, therefore, only the vertical distances (in terms of overshoot and
undershoot) of the stopping point to the target was analyzed.
Analysis of the walking times generally showed that people walked significantly
longer for the successful metaphorical stories (15.7 seconds) than for the unsuccessful
metaphorical stories (12.8 seconds), but that this difference was not reliable in the
nonmetaphorical condition (14.8 and 14.6 seconds). Analysis of the length of
walking (in vertical relationship to the target) again showed that people walked
further for the successful stories (2.4
beyond the object) than for the unsuccessful
ones (2.3
below the object). Analysis of both the thinking and mood ratings
revealed no main effects or interactions for the walk condition. This suggest that
the time and distance effects noted above were not simply attributed to people
© 2006 The Author
Journal compilation © 2006 Blackwell Publishing Ltd
454 R. W. Gibbs

thinking more about one kind of story than the other, or that one story evoked a
different mood which in turn influenced people’s walking behavior (e.g. people
walked less having heard the unsuccessful story because it depressed their mood).
The results of the imagine condition showed that people imagined walking
longer for the successful metaphor stories (11.4 seconds) than for the unsuccessful
metaphoric narratives (9.5 seconds). Unlike the data for the walking condition,
where no difference was obtained, people imagined walking longer in the
unsuccessful condition (12.5 seconds) than in the successful one (9.5 seconds). The
reason for this latter finding is not clear. One might suppose that more difficult
journeys always take longer. But the above results show that this is not the case, at
least when people walk or imagine walking after hearing the metaphorical
narratives. Finally, analysis of the thinking and mood ratings in the imagine
condition indicated no main effects or interactions, which again suggests that the
time effects obtained are not likely attributed to the amount of thought given to
the stories or to participants’ affective responses to the different stories.
These studies suggest that people’s interpretation of the stories partly involved
creating an embodied simulation, or a reenactment, of the relationship journey
alluded to in the different metaphorical narratives. Even though relationships are
not physical entities that literally travel along physical paths, people nonetheless
conceive of relationships in metaphorical ways, especially when prompted to do so
by statements like ‘Your relationship was moving along in a good direction’. This
metaphorical conceptualization is not purely abstract, but embodied in the sense
that participants imagine themselves moving in the different relationship journeys
which subsequently affected their walking, and imagining of walking, as they
thought about the stories.
Of course, the above studies employed a novel methodology, and it is also not
clear whether these findings generalize to other kinds of metaphorical language.
The experiments here also only investigated people’s reflections about the stories
they heard and did not assess their online processing of the different metaphorical
and nonmetaphorical narratives. Nonetheless, the findings are quite consistent
with the idea that reasoning about metaphorical narratives involves the construction
of embodied simulations related to the actions mentioned (e.g. ‘moving along in a
good direction’), a conclusion that is also in line with an emerging literature in
psycholinguistics on the importance of embodiment in nonmetaphorical linguistic
processing (Gibbs, 2006a; Glenberg and Kashak, 2002; Matlock, 2004; Pecher and
Zwaan, 2005), and with claims that the interpretation of fiction is closely tied to
embodied simulation (Oatley, 1999).

5. Conclusion

In her essay ‘The moral necessity of metaphor’, Cynthia Ozick (1986) aptly
commented, ‘Through metaphor, the past has the capacity to imagine us, and we
it… Those who have no pain can imagine those who suffer. Those at the center
© 2006 The Author
Journal compilation © 2006 Blackwell Publishing Ltd
 Metaphor Interpretation as Embodied Simulation 455

can imagine what it is to be outside. The strong can imagine what it is to be weak.
Illuminated lives can imagine the borders of stellar fires. We strangers can imagine
the familiar hearts of strangers’.
The present article describes some empirical research in support of the idea that
metaphor understanding allows us to imaginatively project ourselves into other
people’s minds and worlds. My primary claim here is that this imaginative
engagement arises from metaphor understanding not as an after-the-fact reaction
to metaphor, but as a fundamental part of how we ordinarily interpret metaphorical
meaning. People may create embodied simulations of speakers’ messages that
involve moment-by-moment ‘what must it be like’ processes which make use of
ongoing tactile-kinesthetic experiences. More dramatically, these simulation
processes operate even when people encounter metaphoric language that is abstract,
or refers to actions that are physically impossible to perform. Understanding
abstract events, such as ‘grasping the concept’, is constrained by people’s embodied
experience as if they are immersed in the discourse situation, even when these
events can only be metaphorically realized. This interpretation of the various
findings presented above is congruent with a body of emerging evidence in
cognitive science showing intimate connections between perceptual/sensorimotor
experience and language understanding (Pecher and Zwaan, 2005).
The assorted results of the new studies described in this article do not necessarily
generalize to all kinds of metaphor. Although many metaphoric phrases refer to
bodily activities and sensations, there are other types of metaphoric language that
bear little relation to the human body, or have source domains that are not linked
to embodied experiences. For this reason, embodied activity has not been
demonstrated for all aspects of metaphor comprehension. Nonetheless, bodily
activity provides a major source for metaphorical concepts and the language people
use to refer to these ideas. Simulated body movement may be critical for many
aspects of metaphoric language understanding.
These experiments do not distinguish between the possibility that sensorimotor
activity is actively recruited in metaphor comprehension and the idea that
functionally-independent conceptual representations are activated when linguistic
metaphors referring to abstract concepts are understood. Even if these conceptual
representations for abstract concepts are independent of immediate bodily action,
they still may be partly formed via sensorimotor processes and retain something
about their embodied origins. Under this latter possibility, people’s bodily
experiences of handling physical objects may be used in creating, and maintaining
elaborate conceptual representations for many abstract concepts. But these
‘embodied’ concepts need not be continually tied, and immediately influenced, by
ongoing body activity. As shown here, even imagining appropriate body actions
facilitates processing, which is likely due to activation of relevant pre-motor and
motor cortex regions during mental imagery of the relevant actions.

Department of Psychology
University of California, Santa Cruz
© 2006 The Author
Journal compilation © 2006 Blackwell Publishing Ltd
456 R. W. Gibbs

References

Bergen, B. 2005: Mental simulation in literal and figurative language understanding. In

S. Coulson and B. Lewandowska-Tomaszczyk (eds), The Literal and Nonliteral in
Language and Thought. Berlin: Lang, pp. 255–280.
Buccino, G., Lui, F., Canessa, N., Patteri, I. and Lagravinese, G. 2004: Neural circuits
involved in the recognition of actions performed by non-cospecifics: An fMRI
study. Journal of Cognitive Neuroscience, 16, 1–14.
Currie, G. and Ravenscroft, I. 2002: Recreative Minds. Oxford: Clarendon Press.
Decety, J. and Grezes, J. 1999: Neural mechanisms subserving the perception of human
action. Trends in Cognitive Sciences, 3, 172–178.
Deignan, A. 2005: Metaphor and Corpus Linguistics. Amsterdam: Benjamins.
Feldman, J. and Narayanan, S. 2004: Embodied meaning in a neural theory of language.
Brain & Language, 89, 385–392.
Freeman, D. 1995: Catch(ing) the nearest way: Macbeth and cognitive metaphor.
Journal of Pragmatics, 24, 689–708.
Gallese, V. 2000: The inner sense of action: Agency and motor representations. Journal
of Consciousness Studies, 7, 23–40.
Gallese, V. 2003: The manifold nature of interpersonal relations: The quest for a
common mechanism. Philosophical Transactions of the Royal Society of London, B,
358, 1231–1240.
Gallese, V. and Lakoff, G. 2005: The brain’s concepts’ The role of sensory-motor
system in conceptual knowledge. Cognitive Neuropsychology, 22, 455–479.
Gibbs, R. 1992: What do idioms really mean? Journal of Memory and Language, 31,
485–506.
Gibbs, R. 1994: The Poetics of Mind: Figurative Thought, Language, and Understanding.
New York: Cambridge University Press.
Gibbs, R., Bogdonovich, J., Sykes, J. and Barr, D. 1997: Metaphor in idiom
comprehension. Journal of Memory and Language, 37, 141–154.
Gibbs, R. 2006a: Embodiment and Cognitive Science. New York: Cambridge University
Press.
Gibbs, R. 2006b: Walking the talk while thinking about the talk: Embodied
interpretation of metaphorical narratives. Manuscripts submitted for publication.
Gibbs, R. and O’Brien, J. 1990: Idioms and mental imagery: The metaphorical
motivation for idiomatic meaning. Cognition, 36, 35–68.
Gibbs, R., Strom, L. and Spivey-Knowlton, M. 1997: Conceptual metaphors in mental
imagery for proverbs. Journal of Mental Imagery, 21, 83–110.
Gibbs, R., Lima, P. and Francuzo, E. 2004: Metaphor is grounded in embodied
experience. Journal of Pragmatics, 36, 1189–1210.
Gibbs, R., Gould, J. and Andric, M. in press: Imagining metaphorical actions: Embodied
simulations make the impossible plausible. Imagination, Cognition, & Personality.
Glenberg, A. and Kashak, M. 2002: Grounding language in action. Psychonomic
Bulletin & Review, 9, 558–565.
Glucksberg, S. 2001: Understanding Figurative Language: From Metaphors to Idioms. New
York: Oxford University Press.
© 2006 The Author
Journal compilation © 2006 Blackwell Publishing Ltd
 Metaphor Interpretation as Embodied Simulation 457

Gordon, R. 1986: Folk psychology as simulation. Mind & Language, 1, 158–171.

Grady, J. 1999: A typology of motivation for metaphor: Correlations vs. resemblances.
In R. Gibbs and G. Steen (eds), Metaphor in Cognitive Linguistics. Amsterdam:
Benjamins, pp. 79–100.
Gezes, J. and Decety, J. 2001: Functional anatomy of execution, mental simulation,
observation, and verb generation of action: A meta-analysis. Human Brain Mapping,
12, 1–9.
Harris, P. 1992: From simulation to folk psychology: The case for development.
Mind & Language, 7, 120–144.
Haser, V. 2005: Metaphor, Metonymy, and Experientialist Philosophy: Challenging Cognitive
Semantics. Mouton: Berlin.
Hauk, O., Johnsrude, I. and Pulvermuller, F. 2004: Somatotopic representation of
action words in human motor and premotor cortex. Neuron, 41, 301–307.
Jellema, T., Baker, C., Wicker, B. and Perrett, D. 2000: Neural representation
for the prception of the intentionality of action. Brain & Cognition, 442,
280–302.
Kennedy, J.M. and Vervaeke, J. 1993: Metaphor and knowledge attained from the
body. Philosophical Psychology, 6, 407–412.
Keysar, B., Shen, Y., Glucksberg, S. and Horton, W. 2000: Conventional language:
How metaphoric is it? Journal of Memory and Language, 43, 576–593.
Klatzky, R.L., Pellegrino, J.W., McCloskey, B.P. and Doherty, S. 1989: Can you
squeeze a tomato? The role of motor representations in semantic sensibility
judgments. Journal of Memory and Language, 28, 56–77.
Knuf, G., Aschersleben, G. and Prinz, W. 2001: An analysis of ideomotor action.
Journal of Experimental Psychology: General, 130, 779–798.
Kohler, E., Keysers, C., Umilta, M., Fogassi, L., Gallese, V. and Rizzolatti, G. 2002:
Hearing sounds, understanding actions: Action representation in mirror neurons.
Science, 297, 846–848.
Koski, L., Iacoboni, M., Dubeau, M., Woods, R. and Maziotta, J. 2003: Modulation
of cortical activity during different imitative behaviors. Journal of Neurophysiology,
89, 460–471.
Lakoff, G. and Johnson, M. 1999: Philosophy in the Flesh: The Embodied Mind and its
Challenge to Western Thought. New York: Basic Books.
Lakoff, G. and Turner, M. 1989: More than Cool Reason: A Field Guide to Poetic Metaphor.
Chicago: University of Chicago Press.
Maglio, P. and Matlock, T. 1998: Metaphors we surf by. Workshop on Personal And
Social Navigation in Information Space, Stockholm: Sweden.
Matlock, T. 2004: Fictive motion as simulation. Memory & Cognition, 32, 1389–1400.
McNeil, D. 1992: Hand in Mind. Chicago: University of Chicago Press.
Oatley, K. 1999: Why fiction may be twice as true as fact: Fiction as cognitive and
emotional simulation. Review of General Psychology, 3, 101–117.
Ozick, C. 1986: The moral necessity of metaphor. Harper’s Magazine, May, 64–65.
Ozcaliskan, S. 2003: In a caravanserai with two doors I am walking day and
night: Metaphors of death and life in Turkish. Cognitive Linguistics, 14,
281–320.
© 2006 The Author
Journal compilation © 2006 Blackwell Publishing Ltd
458 R. W. Gibbs

Pecher, D. and Zwaan, R. (eds), 2005: Grounding Cognition: The Role of Perception and
Action in Memory, Language, and Thinking. New York: Cambridge University
Press.
Pulvermuller, F., Haerle, M. and Hummel, F. 2001: Walking or talking?: Behavioral
and neurophysiological correlates of action verb processing. Brain & Language, 78,
143–168.
Richardson, D., Spivey, M., Barsalou, L. and McRae, K. 2003: Spatial representations
activated during real-time comprehension of verbs. Cognitive Science, 27,
767–780.
Roth, W-M. and Lawless, D. 2002: Scientific investigation, metaphoric gestures, and
the emergence of abstract scientific concepts. Learning and Instruction, 12,
285–304.
Umilta, M., Kohler, E., Gallese, V., Fogassi, L. and Fadiga, L. 2001: ‘I know what you
are doing’: A neurophysiological study. Neuron, 32, 91–101.
Wilson, N. and Gibbs, R. 2005: Real and imagined body movement primes metaphor
comprehension. Manuscript submitted for publication.

© 2006 The Author

Journal compilation © 2006 Blackwell Publishing Ltd